Inheritance of Root Dry Matter Content in Sweetpotato

INHERITANCE OF ROOT DRY MATTER CONTENT IN SWEETPOTATO

D. SHUMBUSHA, G. TUSIIME1, R. EDEMA1, P. GIBSON1, E. ADIPALA2 and R.O.M. MWANGA3 Rwanda Agriculture Board (RAB), P. O. Box 5016, Kigali, Rwanda 1College of Agricultural and Environmental Sciences, Department of Agricultural Production, Makerere University, P. O. Box 7062, Kampala, Uganda 2Regional Universities Forum for Capacity Building in Agriculture (RUFORUM), Makerere University, P. O. Box 7062, Kampala, Uganda 3International Potato Center (CIP), P. O. Box 22274, Kampala, Uganda Corresponding author: [email protected]

(Received 26 July, 2013; accepted 9 February, 2014)

ABSTRACT

There has been much emphasis on breeding for increased sweetpotato storage root yield, but less on dry matter yield, and its inheritance. High dry matter content (DMC) is associated with consumer preferences, and is important for the processing industry. This study was conducted to determine the type of gene action controlling DMC and to assess genotype by environment (G x E) interaction effect on DMC in sweetpotato. Five parental clones varying in DMC were hand-crossed in a half-diallel design to generate ten families. Ten genotypes of each family were planted in a trial at Namulonge (swamp and upland environments) and Serere in Uganda in 2009 and 2010. Highly significant (P<0.001) differences were found both between genotypes and between families for DMC. High significant general combining ability (GCA) (P<0.001) and specific combining ability (SCA) (P<0.01) were obtained, meaning that the differences among families for high DMC were due to both GCA and SCA. The relative importance of GCA and SCA was 0.59, indicating that additive gene action was slightly more predominant than non-additive gene action in predicting progeny performance for high DMC. Broad sense heritability (H) estimates for DMC were 0.70 and 0.73, respectively on genotype and family means across environments basis, suggesting that DMC was moderately influenced by the environment. Rapid selection for best genotypes would be possible, since progenies can be predicted from the phenotype of the parents. Parent SPK (GCA = 1.02) was the best combiner. The effect of location was less significant compared to seasons, suggesting the need to evaluate genotypes for several seasons, but in few locations to save resources.

Key Words: General combining ability, half-diallel, Ipomoea batatas, Uganda

RÉSUMÉ

Plusieurs efforts ont été fournis dans le cadre d’augmenter le rendement de la patate douce, mais peu d’efforts visant le rendement sec et son héritabilité. La matière sèche est associée aux préférences des consommateurs et elle est importante dans l’industrie de transformation. Cette étude a été menée dans le but de déterminer le type de l’action génétique contrôlant la matière sèche, ainsi que d’évaluer l’effet d’interaction génotype x environnement sur la matière sèche sur la patate douce. Cinq différents parents en terme de matière sèche ont été croisés en moitié diallèle et dix familles sont obtenues. Les semences sont plantées dans les boites en bois dans les serres à Namulonge, Uganda. Dix génotypes pour chacune des familles sont plantées dans l’essai à Namulonge (environnement marrais et hautes terres) et Serere en bloques complètement randomisées, avec deux répétitions, durant la période Octobre 2009-Mars 2010. Pour déterminer la matière sèche, une quantité de 200 g pour chacune des génotypes a été séchée à 650C jusqu’à ce que le poids soit constant. Les données sont analysées en utilisant le logiciel Genstat. Aptitude générale à la combinaison (AGC) et aptitude spécifique à la combinaison (ASC) sont calculées selon Modèle I, Méthode 4 selon la description de Griffing (1956). Hautes significative (P< 0,001) différences sont trouvées aussi bien entre génotypes que familles pour la matière sèche. Hautes significatives 70 D. SHUMBUSHA et al. AGC (P<0,001) et ASC (P<0,01) sont trouvées, signifiant que les différences observées entre les familles pour la matière sèche sont dues à la fois à AGC et ASC. L’importance relative de l’AGC et ASC était 0,59, ce qui indique que l’action génétique additive était un peu plus importante que l’action génétique non-additive en prédicant la performance des progénitures pour la matière sèche. L’héritabilité en large sens (H) pour la matière sèche était de 0,70 et 0,73 en se basant respectivement sur la moyenne du génotype et celle de la famille sur tous les environnements, suggérant que la matière sèche était modérément influencée par l’environnement. Ceci indique que la sélection rapide des génotypes serait possible, car les progénitures peuvent être prévenues en se basant sur le phénotype des parents. Parent SPK (GCA=1,02) était la meilleure combinant dans cette étude. L’effet des locations était moins significatif comparable aux saisons, suggérant l’importance d’évaluer les génotypes sur plusieurs saisons, mais dans les moins de locations dans le cadre d’économiser les ressources.

Mots Clés: Aptitude générale à la combinaison, moitié diallèle, Ipomoea batatas, Uganda

INTRODUCTION estimated broad-sense heritability of DM content to be 0.97. Sweetpotato (Ipomoea batatas L. (Lam)) ranks Most of the DM in sweetpotato (85 to 90%) seventh in the world, on dry matter basis, among is carbohydrates (Wanda et al., 1987), thus the food crops (FAO, 2008). In sub-Saharan factors that affect the total carbohydrate fraction Africa, it is the third most important crop with are essentially the same as those that influence nearly 90% of the total output coming from eastern DMC. The carbohydrate fraction consists of and southern Africa (Ewell and Mutuura, 1994). starch, sugars, pectin, cellulose and probably Uganda (2,554,000 t), Tanzania (1,379,000 t), and hemicellulose. Sweetpotato starch, which Rwanda (826,000 t) are among the largest contains 79 to 83% amylopectin, is the major sweetpotato producing countries in sub-Saharan carbohydrate, accounting for 65 to 80% of the Africa (FAO, 2008). total dry matter (Muhanna and Rees, 2004). The Most sweetpotato varieties currently concentration of fructose, glucose and total sugar cultivated in SSA have a dry matter (DM) content within sweetpotato roots is negatively correlated that is too low (25-30%) to be used as raw material with root DMC. in the processing industry, which prefers DM The objective of this study was to determine >35% (Lu and Sheng, 1990). Moreover, high DM the inheritance of root DM content in is one of the important attributes that affects sweetpotato, in order to facilitate improvement consumer preferences in most of SSA of the crop for energy value, processing quality (Tumwegamire et al., 2004). For long-term and desirability by farmers. population improvement, a diallel crossing pattern combined with recurrent mass selection is MATERIALS AND METHODS recommended as an effective way to combine favourable genes and alleles in parental The experiment was conducted in a screenhouse genotypes (Huaman and Zhang, 1997). and experimental fields of the National Crops Heritability is a measure of the Resources Research Institute (NaCRRI) at correspondence between phenotypic values and Namulonge and the National Semi-Arid breeding values. High estimate does not explain Agricultural Research Institute (NaSAARI) at how good materials are, only that superior Serere, in Uganda. Namulonge is located at 1,150 parents tend to give the best progeny (Rex, 2002). meters above sea level (masl) and at 0o 32’N, 32o Heritability estimates help to predict the 35’E. It experiences an average high temperature performance of the offspring based on the of 28.4oC, and mean annual bimodal rainfall of performance of their parents, using a particular 1,270 mm (Smit et al., 1997). At Namulonge, two combination of breeding materials and techniques environments were selected for the study, namely of evaluation. Jones (1986) and Courtney (2007), the swamp and upland. respectively, found that DM narrow sense Serere is located in north-eastern Uganda, heritability estimates in sweetpotato, were 0.65 33o27’ E, 1o32’N, and 1,140 masl. The area and 0.92. On the other hand, Wanda et al. (1987) receives annual bi-modal rainfall ranging from 800 Inheritance of root dry matter content in sweetpotato 71 to 1,150 mm. Mean maximum and minimum necessary, until harvest. Roots were harvested temperatures are 33oC and 17oC, respectively five months after planting. One plant at both ends (Kabi et al., 2001). of each row was left as a guard plant, giving a harvest area of 0.72 m2 per sub-plot. Parental material. Five parental sweetpotato clones varying in DMC and other attributes were Data collected. Data were collected on root crossed in a half-diallel design.These five clones characteristics, namely marketable and non- had previously been evaluated for DMC at marketable root number and weight. Roots NaCRRI, with three having high DMC, namely classified as marketable were those over 3 cm in New Kawogo (NKA), SPK004 (SPK) and diameter and without cracks, insect damage or Kyabafuruki (KYA) (Mwanga et al., 2009); and rotten parts (Ekanayake et al., 1990). two having low DMC (Jewel (JEW) and Root characteristics and vine weight were Huarmeyano (HUA)) (Mwanga and Bohac, 2004). evaluated at harvest. Sweetpotato storage root Among the five parents, only JEW was orange- DMC was determined using the method of fleshed. Though self-fertilisation occurs only Benesi et al. (2004). Three undamaged roots from rarely in sweetpotato (Poole, 1955), we each genotype were randomly selected just after emasculated the female parents to eliminate such harvesting. The medial sections of selected roots a possibility. In each of two replications, five for each genotype were sliced thinly, and then clonal plants from each of ten genotypes manually mixed together. A duplicated fresh representing each of the ten full-sib families were sample of 200 g each (w1) was placed in an open- evaluated, totaling 100 genotypes. top paper bag and oven-dried at 65oC for 72 hours to constant weight. The oven-dried samples were Development of F1 plant material. True seeds weighed immediately (w2) and DM (%) was from each selected family were germinated in a calculated gravimetrically. screenhouse. True seeds were scarified by soaking in concentrated (98%) sulfuric acid for Data analysis. Data were subjected to general 30 minutes, to break dormancy imposed by the analysis of variance (ANOVA) using GenStat,12.2 seed coat. Scarified seeds were sown 1 cm deep Edition software. Since there was at least one in sterile soil, in seed-boxes of 150 cm by 90 cm missing genotype recorded in the data from each and spaced 15 cm x 3 cm. location, unbalanced ANOVA was used. Only when significant differences among families were Field experiment. Ten well-established plants, established, was the second step of diallel representing each family, were chosen randomly, analysis performed (Singh and Chaudhary, 1979). and five 30-cm cuttings of each genotype, Fisher’s Protected Least Significance Difference planted in two replications in the field. Experiments test (LSD) at P< 0.05, was used to separate means. were planted at Namulonge, in both upland and The progeny means of DMC were subjected to swamp environments, and at Serere during diallel analyses using GenStat. General combining October – Novemeber 2009 (the second rainy ability (GCA) and specific combining ability (SCA) season) and March of 2010. A split-plot variance components were computed according arrangement, with two replications was used, with to the fixed-effects Model I, Method 4 (only one families arranged as main plots, in a randomised set of F1’s; with parents and reciprocals excluded) complete block design. Genotypes within each as described by Griffing (1956). Correlations family were randomised as sub-plots. Each entry between traits were performed using EXCEL in a replication was represented by a single row programme. containing 5 plants. Parental material was also included in the setups. RESULTS Plant spacing within rows was 30 cm, and 80 cm between rows. Planting and management Dry matter content. There was a very large, practices were similar in all locations, and weeding highly significant interaction of season (S) x was carried out using hand-hoes whenever location (L) (P<0.001) which overshadowed the 72 D. SHUMBUSHA et al. individual main effects of S and of L (Table 1). of all the crosses. The lowest performing family Highly significant differences for DMC were was JEW x HUA with a DMC of 28.0%. The found between genotypes (P<0.001) and families progeny denoted [9] from a cross NKA x HUA, (P<0.001). Genotypes interacted significantly with was the overall best with a DMC of 35.2% (Table all environmental factors (S, L, S x L), but G x L 2). Across seasons, family NKA x SPK was the had smaller interactions than G x S and G x S x L. best performer for DMC (32.7%) (Table 4). There was a significant interaction between Although, the cross Jew x Hua was the lowest in families and environments (F x E (P<0.01) and F x performance in the Serere location, the best S x L (P<0.05)), but no significant interactions performing genotype for DMC (42%) came from were found for F x L and F x S. the same family (data not shown). Both GCA x E and SCA x E were significant. The two crosses with SPK, NKA x SPK (32.5%) Heritability for DMC. Based on variance and SPK x HUA (32.6%), had the highest means components, broad-sense heritability and broad-

TABLE 1. ANOVA for combining ability of DMC in sweetpotato, combined across three locations, two seasons, 2009B and 2010A

Source DF (with missing plots) Mean squares F-test Variance components

Season (S) 1 228.56 0.04ns - Locations (L) 2 5830.30 0.95ns - S x L 2 6118.74 484.84*** - Rep ( R) /(S x L) 6 12.62 3.87*** - Families (F) 9 286.27 4.41*** 1.84 GCA 4 376.94 5.45*** 0.85 SCA 5 202.53 3.29** 1.17 F x Environments (E) 45 31.80 1.85** 0.73 F x S 9 35.64 1.26ns 0.12 F x L 18 33.06 1.17ns 0.12 F x S x L 18 28.21 2.03* 0.72 GCA x E 20 36.01 2.09** 0.31 GCA x S 4 51.66 2.85ns 0.18 GCA x L 8 50.73 3.61* 0.30 GCA xS x L 8 13.48 0.78ns -0.06 SCA x E 25 28.43 1.65* 0.56 SCA x S 5 21.07 0.47ns -0.39 SCA x L 10 20.27 0.50ns -0.51 SCA x S x L 10 40.24 2.34* 1.15 F x [R/(S x L)]=MPE 54 6.25 1.92*** 6.25 Genotypes (G) /F 90 47.39 3.33*** 2.76 G/ F x E 423 14.23 4.37*** 5.48 G/ F x S 89 18.85 1.73** 1.32 G/ F x L 179 14.81 1.36* 0.98 G/F x S x L 155 10.91 3.35*** 3.82 Residual(=SPE) 488 3.26 3.26

Total 1120 36.09

*, **, ***Significant at the 0.05, 0.01 and 0.001 probability levels respectively; ns not significant MPE= Main plot error; SPE= Sub-plot error; SPE= sub-plot error; GCA= variation due to general combining ability; SCA= variation due to specific combining ability Inheritance of root dry matter content in sweetpotato 73 TABLE 2. Within cross variability for sweetpotato dry matter content (%) across environments in Uganda

Genotype Jew x Nka x Hua x Nka x SPK x SPK x SPK x Nka x Nka x Jew x Hua Jew Kya Hua Jew Hua Kya Kya SPK Kya

1 34.5 23.5 33.0 31.9 31.5 33.5 31.3 31.6 32.2 29.7 2 28.3 27.8 29.4 32.1 28.0 31.5 31.1 34.6 32.8 32.0 3 26.8 29.3 31.5 31.9 26.3 30.5 31.1 31.0 31.2 32.6 4 32.3 23.9 28.9 35.0 29.9 31.9 33.3 29.0 35.1 31.7 5 24.4 31.0 29.7 31.9 30.8 34.9 28.9 32.8 30.0 33.6 6 27.9 31.0 32.2 29.8 32.9 33.0 31.1 31.7 31.8 32.5 7 27.2 31.4 30.4 34.8 34.5 34.3 30.8 31.2 31.9 31.4 8 23.5 28.0 30.4 29.4 29.2 31.0 32.7 33.3 32.7 33.3 9 26.6 33.2 29.1 35.2 32.4 32.5 34.3 32.2 34.3 31.1 10 28.6 25.6 32.6 28.7 28.0 32.9 31.4 32.1 33.1 33.4

Mean 28.0 28.5 30.7 32.1 30.3 32.6 31.6 32.0 32.5 32.1 Min 23.5 23.5 28.9 28.7 26.3 30.5 28.9 29.0 30.0 29.7 Max 34.5 33.2 33.0 35.2 34.5 34.9 34.3 34.6 35.1 33.6

Coefficient of variation (CV%) = 6; LSD0.05 = 1.44 sense coefficient of determination (BS-CGD) (P<0.001) negative GCA effects (-1.66). A estimates across environments was 0.70 for combination of low x high DMC parents (JEW x genotype means, and 0.73 for family means, KYA) resulted in a highly significant (P<0.001) respectively. As stated before, only variance positive SCA effect (2.01). A cross between two components for interpreting inheritance were high DMC parents SPK x KYA, produced calculated: progenies with negative SCA effects (-1.13).

2 2 2 Baker’s ratio (2σ GCA /(2σ GCA + σ SCA)) was 0.59. DISCUSSION

Relationship between DMC and some other General and specific combining ability for root variables. Significant differences were found DMC. The significant differences for DMC found between genotypes for marketable root weight among families (Table 1) were due to both GCA (MRW), but not families for this trait (Table 3). and SCA. The magnitudes of the GCA (0.85) and Highly significant interactions were found for G SCA (1.17) variance components suggest that x L, G x S and G x L x S (P<0.001). The means for both additive and non-additive gene action are marketable root weight are shown in Table 4. important in controlling DMC. The relative Significant differences between genotypes importance of GCA and SCA reported here (0.59, (P<0.001) and families (P<0.05) were found for based on Baker’s ratio) suggests that GCA was non-marketable root weight. The interactions slightly more important than SCA in progeny were significant for G x L (P<0.01), G x S (P<0.01) family performance for high DMC. This indicates and G x L x S (P<0.05) for this parameter. Families that additive gene action was slightly more interacted significantly with seasons (F x S, influential than non-additive gene action. This P<0.001), locations (F x L, P<0.05), but not with study conforms to the observations of Sakai the combination S x L. (1964), who reported larger additive than non- additive genetic variance for DMC in General Combining Ability and SCA. GCA and sweetpotato. Cach et al. (2006) also reported quite SCA values and their level of significance are similar results in cassava, where GCA was large presented in Table 5. The high DMC parent SPK and significant for DMC, compared with SCA. showed significant (P<0.01) positive GCA effects These results suggest that, in crossing high by (1.02); while the low DMC parent had significant high DMC parents, progeny should have high 74 D. SHUMBUSHA et al. ) 2 0.13 Across 1 0.1 2010A environments 2009B (both) A sub-plot consisted of 3 middle plants for each genotype (0.72 m † 0.25 0.18 0.22 2009B 2010A = Least Significant Difference. (both) 0.05 1 0.23 0.53 0.38 0.44 0.23 0.25 0.1 0.08 0.18 0.13 0.12 0.08 0.09 0.2 0.1 0.1 2009B 2010A (both) W) means (kg/sub-plot†) for each environment, 2009B and 2010A 2010A 0.69 0.37 0.34 0.18 0.70.50.7 0.40.7 0.1 0.60.24 0.5 0.4 0.3 0.7 0.13 0.1 0.5 0.2 0.1 0.3 0.2 0.1 0.3 0.2 0.2 0.2 0.2 0.2 0.2 0.9 0.4 0.4 0.4 0.3 0.4 0.3 0.6 0.3 0.4 0.3 0.3 0.5 0.4 0.3 0.4 0.2 0.4 0.3 0.3 0.4 0.4 0.3 0.80.80.60.5 0.11.0 0.5 0.2 0.4 0.5 0.2 0.7 0.4 0.5 0.3 0.6 0.2 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2 0.1 0.2 0.2 0.1 0.6 0.1 0.3 0.3 0.5 0.2 0.4 0.2 0.2 0.4 0.2 0.2 0.2 0.2 0.4 0.6 0.3 0.4 0.3 0.3 0.2 0.5 0.3 0.2 0.3 0.4 0.2 0.4 0.3 0.3 0.3 0.50.7 0.3 0.7 0.4 0.7 0.1 0.3 0.1 0.4 0.1 0.3 0.3 0.1 0.2 0.3 0.3 0.2 0.3 0.3 0.2 0.5 0.3 0.4 32.4 39.3 62.5 35.2 45.6 54.4 24.4 29.9 28.6 Serere Swamp Upland Combined A A A A x KY X KY 0.05 ABLE 3. Marketable root weight (MR 2009B CV = coefficient of variation; s.e standard error; s.e.d error the difference; LSD T LSD Family sed NKA X SPK JEW X KY G. mean CV% s.e NKA x HUA SPK X JEW SPK x HUA SPK X KY NKA JEW x HUA NKA x JEW HUA Inheritance of root dry matter content in sweetpotato 75 Across seasons

2010A 2009B Both 2010A 2009B Both at Serere and Namulonge (swamp upland) in Uganda A 2009B 2010A Both = Least significant difference 005 2010A 2.94.1 3.0 5.5 2.9 8.2 3.4 6.8 2.9 6.2 3.1 4.6 2.6 5.9 2.7 5.7 2.1 5.8 37.637.136.2 28.7 29.536.2 30.8 33.2 33.3 28.1 33.5 24.0 24.3 32.1 25.7 30.5 32.5 23.0 31.4 27.3 30.6 28.4 28.5 33.1 34.1 26.8 34.5 36.9 38.7 32.8 33.7 35.0 36.4 36.1 34.1 31.6 34.5 31.8 32.1 32.1 33.6 30.7 32.0 32.0 31.6 32.7 32.2 31.1 34.935.4 24.7 27.1 29.836.4 31.3 21.1 26.5 20.3 26.4 31.5 26.3 26.1 23.7 23.3 28.0 31.6 30.4 32.7 28.8 34.0 30.3 33.5 32.2 36.5 28.0 28.7 35.0 27.9 29.1 32.0 28.0 28.8 31.5 31.7 35.737.335.0 27.036.6 29.4 26.6 31.4 30.5 33.3 30.8 23.5 33.5 22.0 20.3 30.3 22.9 32.8 31.3 32.8 26.9 27.4 25.8 32.7 27.9 34.0 33.8 35.4 34.3 37.0 37.5 34.1 38.5 35.5 35.6 36.4 30.6 31.1 29.7 30.9 31.2 33.0 31.8 34.0 30.7 32.1 30.6 32.6 Serere Swamp Upland Seasonal means A A A A x KY X KY 0.05 ABLE 4. Diallel mean for dry matter content (%) each season B and NKA X SPK JEW X KY Grand mean LSD CV% c.v = Coefficient of variation, LSD JEW x HUA NKA x JEW HUA NKA NKA x HUA SPK X JEW SPK x HUA SPK X KY Family 2009B T 76 D. SHUMBUSHA et al. TABLE 5. Estimates of GCA effects (in bold) and SCA (off diagonal) effects for sweetpotato DMC across environments in Uganda

Parent NKA SPK JEW HUA KYA

New Kawogo (NKA) 0.32ns 0.18ns -1.03ns 1.01ns -0.17ns Kakamega (SPK) 1.02** 0.14ns 0.82ns -1.13* Jewel (JEW) -1.66*** -1.12* 2.01*** Huarmeyano (HUA) -0.36ns -0.71ns Kyabafuruki (KYA) 0.68ns

*, **, *** Significant at the 0.05, 0.01 and 0.001 probability levels respectively; ns not significant. Standard error, GCA = 0.38; SCA = 0.52

DMC values with means close to the mid-parent two parents with low DMC. Furthermore, cross value. JEW x KYA was the best performer for MRW, with a combined mean of 0.4 kg sub-plot-1. Dry matter content (DMC) genotype. Root DMC Orange-fleshed sweetpotatoes are important values for family means ranged from 23.5% for in addressing vitamin A deficiency in developing both the JEW x HUA and NKA x JEW families to countries, and are being promoted especially in 35.2% for NKA x HUA (Table 2). These values Eastern Africa (Tumwegamire et al., 2004). Jewel, are within the ranges of DMC for the current therefore, could be a good parent for obtaining sweetpotato genotypes in Eastern Africa, as progeny with orange flesh as well as moderate to reported by Brabet et al. (1998). The results show high DMC. that the two low DMC parents, JEW and HUA, The coefficient of variation for MRW was contributed negatively to the lowest DMC generally high, especially for the swamp 2009 B genotype mean (23.5%), among all crosses. season (62.5%) and the upland 2010 season Highly significant differences for DMC were (54.4%) (Table 3), indicating large plot-to-plot found between both genotypes within families variability (Table 3). This agrees with the and between families. These results concur with statement of Andrade et al. (2009) that those of Kanju (2000) who studied sweetpotato sweetpotato experimental trials have a very large in South Africa. The variance component for plot error for yield traits. The same authors have genotypes within families was higher (2.76) than reported cv values of 35-45% in plots with 15 that for families (1.84), indicating that selection plants, and cv values of 25% in plots with more of specific genotypes within families for high than 60 plants. This shows that the cv plot error DMC is very important. can be improved through using relatively large Two families involving the high DMC parent samples of plants.This indicates that plot to plot SPK had the highest means of all the crosses variability was very high for this trait. In some (NKA x SPK with 32.7%, SPK x HUA with 32.6%). cases, there were no roots at all, or only non- Genotype 9 from a high by low DMC cross, NKA marketable roots were present x HUA (Table 2), was the best genotype averaged across the six environments, with a mean of Performance of genotypes. Differences were not 35.2%; followed by genotype 4 from family NKA large among family means for DMC averaged over x SPK (35.1%). Based on the average performance locations and seasons (Table 4). Namulonge of genotypes within families, the lowest swamp gave the lowest mean for DMC (26.8%); performing genotype overall was from the family while the highest locational mean was Namulonge NKA x JEW, with a DMC mean of 23.5% (Table upland (34.5%). Seasons, 2009B and 2010A 2). resulted in nearly equal DMC means of 30.7 and Genotype 1 from JEW x HUA, a cross from a 31.6%, respectively, although there were strong low by low parental DMC, ranked seventh in interactions of seasons with locations performance (34.5%), indicating that it is possible (environments). Parent SPK was involved in two to get good genotypes from crosses involving Inheritance of root dry matter content in sweetpotato 77 of the three best families for DMC, which agreed here suggest that rapid selection would be with the GCA value for this parent (Table 4). possible, since clonal reproduction allows for the preservation and utilisation of superior General and specific combining ability effects genotypes, regardless of the relative contributions for DMC. Based on GCA values (Table 5), the of additive or non-additive types of gene action. high DMC parent SPK had the highest positive and significant GCA effects (GCA = 1.02, P<0.01). ACKNOWLEDGEMENT However, its specific cross with another high DMC parent, KYA, produced progenies with We are greatly indebted to the Alliance for a significant negative SCA effects (SCA = -1.13, Green Revolution in Africa (AGRA) for providing P<0.05). This implies that the non-additive gene financial support for this study, the National action arising from the combination of parents, Crops Resources Research Institute, Namulonge SPK and KYA, resulted in the cross performing (NaCRRI) for adequately supplying all the below the expectation based on GCA effects. The necessary facilities that enabled the first author reason for this is not clear, though since run experiments promptly. Our acknowledgement sweetpotato is highly heterozygous and also goes to the Rwanda Agriculture Board hexaploid, favourable gene combinations can be (RAB), the former “Institut des Sciences broken down in progeny from a cross of parents Agronomiques du Rwanda (ISAR)” for giving selected for a desirable trait. the senior author study leave. Dr. Gorrettie The orange-fleshed root parent, JEW, had Ssemakula, Dr. Yona Baguma, Charles Niringiye, negative and high GCA effects (GCA = -1.66, Eugene Ekinyu, Yakub Sekamwa and Moses P<0.001) (Table 5). Though this parent had Mwondha are also acknowledged for their negative GCA effects, it produced progeny with support. a positive and highly significant SCA when crossed with parent KYA (SCA = 2.01, P<0.001). REFERENCES Similar results have been reported by Chiona (2009) on β-carotene content in sweetpotato in Andrade, M., Barker, I., Cole, D., Dapaah, H., Zambia. These results imply that parents should Elliott, H., Fuentes, S., Grüneberg, W., not be discarded based on their negative GCA Kapinga, R., Kroschel, J., Labarta, R., Lemaga, effects, since they can express their performance B., Loechl, C., Low, J., Lynam, J., Mwanga, R., in specific crosses and in specific genotypes. Ortiz, O., Oswald, A. and Thiele, G. 2009. Unleashing the potential of sweetpotato in Estimates of heritability. Since genotypes were sub-saharan Africa: Current challenges and random representatives of families, calculation way forward. International Potato Center of heritability estimates is appropriate, and the (CIP), Lima, Peru. Working Paper 2009-1. broad-sense heritability estimate for DMC was 197pp. 0.70 for genotype means across environments. Benesi, I.R.M., Labuschagne, M.T., Dixon, A.G.O. With regard to families, parents were considered and Mahungu, N.M. 2004. Genotype x fixed, implying that it is appropriate to report the environment interaction effects on native broad-sense coefficient of genetic determination cassava starch quality and potential for starch (BS-CGD), which was 0.73 for family means across use in the commercial sector. African Crop environments. These estimates are relatively high Science Journal 12: 205-216. compared to the range 0 to 1, suggesting that the Brabet, C., Reynoso, D., Dufour, D., Mestress, trait was only moderately influenced by the C., Arredondo, J. and Scott, J. 1998. Starch environment. This implies that rapid selection content and properties of 106 sweetpotato would be possible since good progeny can be clones from the world germplasm collection predicted from the phenotype of the parents. held at CIP, Peru. CIP Program Report 1997- Jones (1980) and Chiona (2009) reported nearly 1998. equal estimates for narrow sense heritability (h2) Cach, N.T., Lenis, J.I., Perez, J.C., Morante, N., for root DMC. The heritability estimates reported Calle, F. and Ceballos, H. 2006. Inheritance of 78 D. SHUMBUSHA et al. useful traits in cassava grown in sub-humid sweetpotato rooting characteristics on conditions. Plant Breeding 125: 177-182. infestation and damage by Cyllas spp. Chiona, M. 2009. Towards enhancement of B- African Crop Science Journal 9:165-174. carotene content of high dry mass Lu, G.Q. and Sheng, J. L. 1990. Application of sweetpotato genotypes in Zambia. PhD near infrared reflectance spectroscopy (NIRS) Thesis, University of KwaZulu-Natal, in sweet potato quality breeding. Scientia Pietermartizburg, South Africa. Agricultura Sinica 23:76-81. Collins, W., Wilson, G., Arrendel, S. and Dickey, Mwanga, R.O.M. and Bohac, J. 2004. L. 1987. Genotype x environment interactions Development of high yielding multiple in sweetpotato yield and quality factors. resistant sweetpotato germplasm. In: Journal Amer. Soc. Hort. Science 112: 579- McKnight Foundation Collaborative 583 Research Program Report. Cornell, USA. Courtney, M. 2007. Genotypic variability and http://mcknight.ccrp.cornell.edu (Accessed inheritance of iron and zinc in sweetpotato. 15 September 2010). Louisiana State University, Baton Rouge, Mwanga, R.O.M., Odongo, B., Niringiye, C., MSc Thesis, USA. Alajo, A., Kigozi, B., Makumbi, R., Lugwana, Ekanayake, I.J; Malagamba, P. and Midmore, D.J. E., Namakula, J. and Mpembe, I. 2009. 1990.Effect of water stress on yield indices of NASPOT 7, NASPOT 8, “NASPOT 9 O”, sweetpotatoes. In: Howeler, R.H.(Ed.). “NASPOT 10 O” and “Dimbuka-Bukulula” Proceedings 8th Symposium of the sweetpotato. HortScience 44: 828-832. International Society for Tropical Root Crops. Poole, C.F. 1955. Sweetpotato genetic studies. Bangkok, Thailand. 724pp. Hawaii Agricultural Experiment Station. Ewell, P.T. and Mutuura, J. 1994. Sweetpotato in Technical Bulletin No 27. the food systems of eastern and southern Rex, B. 2002. Breeding for quantitative traits in Africa. In: Tropical root crops in a plants. Stemma Press, Minnesota, USA. developing economy. Proceedings of the Sakai, K. 1964. Studies on the enlargement of 9thSymposium of the International Society of varieties and the improvement of selection Tropical Root Crops: 20-26 October 1991, methods in sweetpotato breeding. Bulletin Accra, Ghana. of Kyushu Agricultural Experiment Station Food and Agricultural Organisation of the United 9: 247- 397. Nations (FAO). 2008. FAO Statistics, Rome, Smit, N.E.J.M., Downham, M.C.A., Odongo, B., Italy. Hall, D.R. and Laboke, P.O. 1997. Griffing, B. 1956. Concept of general and specific Development of pheromone traps for control combining ability in relation to diallel crossing and monitoring of sweetpotato weevils, systems. Australian J. Biol. Sci. 9: 463-493. Cylas puncticollis and C. brunneus in Huaman, Z. and Zhang, D.P. 1997. Sweetpotato, Uganda. Kluwer Academic Publishers. in biodiversity in trust: Conservation and use Entomologia Experimentalis et Applicata of plant genetic resources in CGIAR Centres, 85: 95-104. Cambridge University Press, United Singh, R.K. and Chaudhary, B.D. 1979. Biometrical Kingdom. methods in quantitative genetic analysis. International Potato Center (CIP). 2008. Kalyani Publishers, New Delhi, India. Procedures for the evaluation and analysis Tumwegamire, S., Kapinga, R., Zhang, D., of sweetpotato trials. 22pp. Crissman, C. and Agili, S. 2004. Opportunities Jones, A.1986. Sweetpotato heritability estimates for promoting orange-fleshed sweetpotato as and their use in breeding. HortiScience 21: a mechanism for combat vitamin A-deficiency 14-17. in sub-saharan Africa. Africa Crop Science Kabi, S., Ogenga-Latigo, M.W., Smit, N.E.J.M., Journal 12: 241-252. Stathers, T.E. and Rees, D. 2001. Influence of